The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill fetches a JSON schema from https://api.kestra.io/v1/plugins/schemas/flow. This is a legitimate operation used to ground the AI's generation in the current platform's specifications. The URL belongs to the skill author's official domain.
[INDIRECT_PROMPT_INJECTION]: The skill ingests external data from the Kestra API to guide the generation of YAML files. * Ingestion points: The curl command in SKILL.md fetches remote schema data. * Boundary markers: The instructions explicitly tell the agent to 'Read the raw JSON output to validate every type' and 'do not pipe it through any interpreter'. * Capability inventory: The skill uses curl for data retrieval and provides logic for YAML generation. * Sanitization: The skill contains a fallback error message if the schema cannot fulfill a user request, preventing the generation of invalid or potentially harmful structures.

kestra-flow